Re: [RFC] x86: xsave/xrstor support, ucontext_t extensions

From: Mikael Pettersson
Date: Thu May 22 2008 - 18:17:32 EST

Chris Wright writes:
> * H. Peter Anvin (hpa@xxxxxxxxx) wrote:
> > Suresh Siddha wrote:
> >>
> >> hpa, What is the virtualization problem? Are you referring to perf problem?
> >> As you noted, regular non-rt signal handlers won't need this cpuid check. It's
> >> needed only for those who manually look at non-rt signal frames and interpret it.
> >> And also, they can do this check only once and not everytime.
> >>
> >
> > No, relying on CPUID and vdso both have implications for virtualization.
> hmm, what implications? cpuid is dealt with already.

By whom?

Case in point: only a few weeks ago, we (as in LKML hackers) tried
to debug a user's bug report. Something very strange seemed to be
going on. It turned out the user was actually using a virtualisation
engine (qemu, not the latest) which presented highly inconsistent
cpuid data to the kernel, causing major confusion.

So I agree with hpa. The kernel has decided to use a particular
sigframe layout. It should communicate this to user-space in a robust
manner that is independent of how the kernel arrived at this decision.
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at